The non-Hodgkin's lymphomas (NHL) represent a major cause of cancer-related morbidity and death. There are approximately 56,000 new cases of non-Hodgkin's lymphoma each year in the U.S., resulting in over 25,000 deaths annually. Approximately 10-20% of NHL cases involve the brain. CNS involvement of NHL is associated with an adverse prognosis and can occur by two distinct pathways: primary CNS lymphoma and secondary dissemination of systemic lymphoma to the brain. Beyond the fact that the majority of these tumors express the cell surface molecule CD20, there is an extreme paucity of molecular information regarding the biological basis of CNS involvement of NHL. New treatments are needed to treat these complications. The goal of this proposal is to begin to use targeted therapies to treat CNS lymphomas and to define the biological basis for drug resistance and possibly CNS tropism in these tumors. Ultimately, insights into the molecular features which underlie CNS involvement may lead to the development of more rational therapeutics to treat or to prevent these complications. This proposal has three specific aims: One, to perform a phase I clinical trial which investigates the safety and potentially the efficacy of intrathecal administration of the anti-CD20 antibody, rituximab, in patients with recurrent lymphomatous meningitis. Two, to use this trial a mechanism for the procurement of lymphomatous meningitis specimens for gene expression profile analysis to determine the mechanistic basis for dissemination of lymphoma cells within the leptomeninges and the basis for drug resistance which commonly is manifest within the leptomeningeal compartment. Three, to identify prognostic markers in primary CNS lymphoma. Our experience suggests that there are at least two subtypes of primary CNS lymphoma, one which is sensitive to methotrexate-based therapy and which is associated with prolonged survival, and one which manifests drug-resistance and is associated with early progression. We will perform gene expression profile analysis of these tumors from diagnostic specimens obtained from a frozen tumor bank to identify a set of genes, which can be used to predict outcome.